极小子样下参数随机分布的电主轴可靠性分析

Reliability Analysis of Motorized Spindle with Parameters Randomly Distributed under Extremely Small-scale Samples

对某型号电主轴进行了极小子样下的可靠性分析,将虚拟增广法和Boot-strap法结合,对原始试验样本虚拟扩展,建立了参数服从随机分布的退化模型和电主轴可靠度函数,并利用Bayes验前分布,对该随机系数下电主轴可靠度函数进行修正,得到修正后的可靠度函数。将修正后随机系数下电主轴可靠度函数曲线与伪寿命下得到的可靠度函数曲线进行对比。结果表明,随机系数法得到的可靠度函数具有更高的精度,同时也避免了将模型参数看作固定值即采用伪寿命法进行可靠性评估时产生的误差。进一步说明根据极小子样试验数据可以通过该方法得到同类产品的退化特征,具有一定的适用性。

For the reliability analysis of a type of motorized spindle under the case of extremely-small samples,the virtual augmented method and Boot-strap method were used to virtually extend the original sample,then the degradation model and the reliability function of which the parameters followed random distribution were established.Besides,Through Bayes prior distribution,the reliability function under random coefficients was modified.The reliability function of the motorized spindle under random coefficients after modification was compared with the reliability function deduced under the case of pseudo-life.The results showed that the reliability function obtained by the random coefficient method had a higher precision,and the errors generated by the reliability evaluation using pseudo-life method in which the model parameters were regarded as fixed values were avoided.The method proposed in this paper can be used to obtain the degradation characteristics for the similar products with certain applicability according to the extremely-small sample experimental data.